1704
EDWARD R. ATKINSON AND H. J. LAWLER
Vol. 62
From the fraction, b. p. 180-190” (2 mm.); n% 1.6355, drolysis, gave 1,2,4-triphenylbutene-(2)(XVII) as a a crystalline crop was secured on trituration with light- colorless, viscous oil, b. p. 140” (0.03mm.); #” 1.5992. petroleum (b. p. 80-lOO0). After recrystallization from Anal. Calcd. for C~2H20: C, 93.0; H , 7.0. Found: C, isopropyl alcohol, it formed hexagonal plates, m. p. 110”, 92.9;H, 7.0. and was identical with the hydrocarbon, described by Summary Bergmann and Ukai’ (m. p. 104-105’). With 2 mols of picric acid in isopropyl alcohol, orange-red needles are 1. The different reactivities of 1- and 2-vinylformed, which on standing turn into a mixture of dark-red naphthalenes and 2-vinylanthracenes toward macrystals and the colorless plates of the initial hydrocarbon; leic anhydride are examined. 2-Isopropenylanon further standing, the dark-red crystals disappear too. On heating, the hydrocarbon also dissolves, and when the thracene shows addition exclusively a t the 9,lOsolution cools down to room temperature, the same effect double bond. is reproduced. 2. The reactivity of 9-vinylphenanthrenes is The liquid isomer, remaining after the separation of the compared with the conjugated system of the corsolid hydrocarbon, was used for the following reactions. responding a-vinylstilbenes, as regarding the Reaction with Maleic Anhydride.-The hydrocarbon (1 g,) and maleic anhydride (0.4 g.) in xylene (10 cc.) diene reaction and the addition of alkali metals, were boiled for three hours. On addition of low-boiling and the ambiguous character of the 9,10-double light-petroleum, 3,4,6-triphenyl-1,2,3,6-tetrahydrophthalicbond of phenanthrene is demonstrated. anhydride (XIII) crystallized. From butyl acetate, 3. 9-Allylphenanthrene, too, shows the peculineedles, m. p. 208-209”; yield 0.3 g. Anal. Calcd. for arities of the 9,lO-double bond of phenanthrene C2BH2003: C, 82.1;H, 5.3. Found: C, 82.6; H, 5.8. Reaction with Lithium Metal.-The reaction was by the two ways it reacts with lithium. carried out as above; the dark-brown solution, on hy- REHOVOTH, PALESTINE RECEIVED OCTOBER 17, 1939
[CONTRIBUTION FROM
THE
CHEMICAL LABORATORIES OF THE UNIVERSITY OF NEW HAMPSHIRE A N D TRINITY COLLEGE (CONN.)]
A Direct Synthesis of Resolvable Biaryls’ BY EDWARD R. ATKINSONAND H. J. LAWLER It has long been known that diazotized amines decompose in the presence of certain reducing agents to form biaryls along with other products, including the common replacement derivatives, as shown by the generalized equation RNZX +R-R H R-N-R
+ R-N=N-R + R-0-R + + R X + R H + ROH + RNHi + Ns
The reaction, which occurs in aqueous solution, invariably leads to symmetrically substituted biaryls2 and seems to be a genuine reduction; in these respects it differs from the better known Gomberg reaction3 which is currently being investigated by Waters and others4 The facts concerning the reaction in question may be correlated with respect to (a) the nature of the diazotized amine, (b) the reducing agent used and (c) the experimental conditions which (1) A preliminary notice of this work was presented at the Rochester meeting of the A. C. S., September, 1937. (2) The possibility of preparing unsymmetrical biaryls by the use of mixtures of diazotized amines is being investigated. (3) Gomberg and Bachmann, THISJOURNAL, 46, 2339 (1924); Gomberg and Pernert, ibid., 48, 1372 (1926). (4) Hey and Waters, Chem. Rev., 21, 175-183 (1937), and later papers in the Journal of the Chemical Society by Waters, Grieve, Hay, Heilbron.
include a number of variables. I n an effort to determine the relative importance of these factors and to extend the general applicability of this reaction to matters of synthesis, an extensive study of the reaction has been undertaken. At the present time i t is not possible to state the potential use of the reaction in the synthesis of biaryls and because of this it has not had wide application. However, Vorlander and M e ~ e r , ~ and Huntress and his students’ have used this method for the preparation of diphenic acid and a number of symmetrical substitution products; the preparation of certain binaphthyl dicarboxylic acids from diazotized naphthylamine carboxylic acids has been described in the patent literature. We have prepared d,l-4,G,4’,G‘-tetrachlorodiphenic acid (11) and the analogous bromo compound (111) in two steps, starting with the readily available anthranilic acid. (5) Vorliinder and Meyer, A n n . , 820, 122-144 (1902). (6) Hunn, THISJOURNAL, 46, 1024-1030 (1923). (7) (a) Huntress, “Organic Syntheses,” John Wiley and Sons, New 66, York, 1932, Collective Volume I, page 216; (b) THIS JOURNAL, 2561 (1933); (c) 66, 4268 and 4270 (1933). (8) German Patent 445,390, Friedlander, 16, 300 (1928).
DIRECT SYNTHESIS OF RESOLVABLE BIARYLS
July, 19.10
1705
Experimental Part All melting points reported in this paper were taken by the method described in Mulliien's "Identification of Pure Organic Compounds," Vol. I, page 218, on a 360" rod form melting point thermometer immersed in the sulfuric acid VCOOH . ClUCOOH to the 0" point. All melting points are uncorrected. C1))COOH NHn "* 3,5-Dichloro-2-aminobenzoicAcid (I) .-Two alternative methods were used for the preparation of this substance. I Method 1.-In a typical run based on the method dec1 scribed by Eller and Klemm,lz 25 g. (0.18 mole) of anthraI1 nilic acid was dissolved in 100 cc. of thiophene-free benzene and a solution of 70 g. (0.52 mole) of sulfuryl chloride in 50 cc. of benzene added slowly with cooling and vigorous shaking. The resulting suspension was heated on a waterBr COOH bath for four hours, cooled, and diluted with one volume of ligroin (b. p. 90-120'). After purification and recrystallization from benzene-alcohol, the light brown product BrflCOOH had m. p. 227-230", weight 19 g. (51% yield); this material was satisfactory for conversion into the tetrachloroBr diphenic acid. I11 Method 2.-Direct chlorination of from 10 to 30 g. of anthranilic acid dissolved in glacial acetic acid as described The halogenation step was thoroughly examined by Elion'3 gave products which could be purified by leachand modifications of published procedures made ing with boiling benzene until the process no longer reso as to obtain optimum yields. In the coupling moved a red tarry substance. The product had m. p. step by using hydrated cuprous oxide and a set 223-228"; weight 8.9 g. (59% yield); it was satisfactory of experimental conditions, the development of for conversion into the tetrachlorodiphenic acid. Material prepared by either of the above methods could which will be described in a subsequent paper, we be further purified by recrystallization from benzenehave obtained yields of 50% for the chloro com- alcohol; white or cream needles, m . p. 230-231 ', ElionIs pound and 37yofor the bromo compound. Easy records m. p. 231 '. d,l-4,6,4',6'-Tetrachlorodiphenic Acid (II).-The folrecovery of dihalogenoanthranilic acid (I) from the by-productsg led to net yields of 61 and 53yo, lowing procedure is that of a typical run. Runs using onehalf the amounts described also have been carried out. respectively. Both the tetrachloro- and tetraTen grams (0.048 mole) of 3,5-dichloro-2-aminobenzoic bromodiphenic acids were resolved into their ac- acid was dissolved in a mixture of 100 cc. of water and 26 cc. tive forms by brucine, although no extensive study of 2 N sodium hydroxide (0.052 mole). To this was of optimum resolution procedures was made. added 4 g. (0.056 mole) of sodium nitrite. Mechanical These active forms are quite stable to racemiza- stirring was often necessary to dissolve the latter, since the solution is almost saturated with respect to the sodium tion. We now have then a convenient and rapid salt of the acid. This solution was added in 5-cc. portions synthesis of optically active diphenic acids. The with vigorous shaking to a solution of 40 cc. of 6 N hydrouse of this class of substances as resolving agents chloric acid in 20 cc. of water, the latter solution being kept at 0'. The diazotization took place rapidly and the rehas been discussed by Ingersoll and Little.'O Yuan and Hsiil' have previously described a sulting solution was kept a t 0" in the dark for about onehalf hour, then filtered by suction immediately before use tetrabromodiphenic acid prepared by the usual if not entirely clear. Ullmann method from the methyl ester of 3,5-diJ. T. Baker A . R. grade of cuprous chloride was purified bromo-2-iodobenzoic acid. Their acid should be by washing with 1:20 sulfuric acid, then with glacial acetic identical with ows, yet there are very few points acid. It was dried a t 100" and stored in the dark. Twentyof similarity; even the relative solubility of the four grams (0.242 mole) of this pure cuprous chloride was suspended in 200 cc. of water in a 500-cc. round-bottom two alkaloidal salts is reversed (the same solvent flask. Forty-two cc. of 6 N sodium hydroxide (0.252 being used). For this reason the rather unex- mole) was added a t about one drop per second while the pected easy racemization of the active forms de- contents of the flask were vigorously stirred. The rescribed by these workers may not be significant. sulting orange hydrated cuprous oxide was washed twice As stated above, our active acid was not easily with water by decantation, then filtered by suction and sucked as dry as possible. It was transferred mechanically racemized. to the flask in which the coupling reaction was to be carried (9) This material has been shown to be a true by-product and not out.
v
0 I
v
to result from incomplete diazotization. (IO) Ingersoll and Little, THISJOURNAL, 56, 2123 (1934). (11) Yuan and Hsu, J . Chinrsr Chem. S a . , 3, 206 (1935).
(12) Eller and Klemm. Bcr., 66, 222 (1922) (13) Elion, Rec. traw. chim.. 44,1106 (1926).
1706
EDWARD R. ATKINSONAND H. J. LAWLER
Vol. 62
The coupling reaction was carried out in a 500-cc. threeFour grams (0.01052 mole) of the acid dissolved in 50 necked round-bottom flask immersed in a n ice-bath. The cc. of boiling ethyl acetate was added quickly to a boiling use of a creased flask described by Morton14 is not recom- solution of 8.8 g. (0.0223 mole) of anhydrous brucine in mended. The reducing agent was suspended in a mixture 200 cc. of ethyl acetate. A heavy crystalline precipitate of 30 cc. of ammonium hydroxide (28%) and 40 cc. of appeared. Successive concentrations of the filtrate led water. During the reaction a total of 16 cc. of ammonium to the recovery of 11 g. of salts. Since optical examinahydroxide and 10 cc. of water additional were added from tion of the fractions showed them to be mixtures, they were time t o time in small portions. The diazonium solution subjected to a simple fractional crystallization from ethyl was added at about one drop per second through an alcohol. addition tube which dipped below the surface of the reducThe less soluble salt (ZAJB) had m. p. 264-265". ing suspension and then bent upward. The end of this Rotation. 0.0559 g. made up to 1.95 cc. in chloroform a t tube was constricted t o a diameter of one millimeter. This 24" gave CYD -0.76; 1, 1; [ C Y ] Z ~ D -26.5'. device prevented premature reaction of the entering diAnal. Calcd. for CipHeOaCla~(CasHie0aNz)z:C1, 12.16; azonium solution with the vapors of the ammoniacal N, 4.80. Found: C1, 12.12, 12.12; N, 4.77; 4.84. reducing solution. Motor stirring was used and was quite The crystalline fraction of the more soluble salt (dAdB) vigorous in the vicinity of the addition tube. Troublesome having the lowest specific rotation had m. p. 254-259 ". foaming caused by the vigorous evolution of nitrogen was Rotation. 0.0519 g. made up to 1.95 cc. in chloroform a t prevented with a few drops of n-butyl alcohol. At the 24Ogave CYD -0.21; I , 1; [cY]**D -7.9". conclusion of addition, excess hydrated cuprous oxide was Anal. Calcd. for C1aH~04C14.(CzsHze04Nz)a: N, 4.80. always present. The solution was brought to a boil and Found: N, 4.61. acidified slowly with concentrated hydrochloric acid until During another resolution experiment using aqueous alacid to congo red; rapid acidification usually gave a tarry cohol as solvent and one-half the equivalent amount of product; 50 or 60 cc. additional acid was then added and the crude product which had precipitated was allowed to brucine, there was isolated the acid brucine salt of the d stand overnight. The crude was filtered off, leached with a acid, m. p. 263-265" decompn. after drying a t 110". Rotation. 0.0713 g. made up to 1.95 cc. in chloroform a t cold saturated solution of ammonium chloride at its boiling 25Ogave CYD-0.56; 1, 1; [ C Y ] ' ~ D-15.3". point to remove copper salts, filtered from the hot solution Anal. Calcd. for C~4He04C14~CzsHzs04Nz: C1, 18.3G; and washed free of chloride. After drying it had m. p. N, 3.62. Found: C1, 18.1; N, 3.6, 3.7. 177-233', weight 9 g. 1-4,6,4',6'-TetrachlorodiphenicAcid.-Two grams of the The crude was heated to 150' with 40 cc. of concentrated sulfuric acid and then allowed to cool; if more sulfuric acid less soluble salt was decomposed with dilute sodium hywas used the entire crude dissolved and the tetrachlorodi- droxide and the liberated brucine removed. The solution phenic acid crystallized out on cooling; normally not all was acidified with 6 N sulfuric acid and the precipitated the solid dissolved. The insoluble material was filtered Z-acid recrystallized twice from ethyl acetate-ligroin; m. p. by suction from the sulfuric acid, using a glass filter; 240-256". Rotation. 0.0369 g. made up to 1.95 cc. in chloroform a t it was washed on the filter with cold concentrated sulfuric D acid until white, the washings being added to the filtrate 25" gave (YD -2.44; I , 1; [ C X ] ~ ~-129'. (the examination of which is described below). The white Anal. Calcd. for ClaHeO4Cln: C1, 37.4. Found: C1, solid was then boiled with water to remove residual sulfuric 36.9. acid and finally filtered and dried; weight, 4.5 g. (49% d4,6,4',6'-Tetrachlorodiphenic Acid.-One gram of the yield). I n some cases the product was recrystallized more soluble salt was decomposed as above giving an acid, from ethyl acetate-ligroin. Combinations of alcohol or rn. p. 233-250'. acetone with water were not suitable because the product Rotation. 0.0308 g. made up to 1.95 cc. in chloroform invariably separated as a n oil. d,l-4,6,4',6'-Tetrachloro- at 25" gave CYD+1.99; 1, 1; [CY]% +126". diphenic acid (11)is a white crystalline solid, m. p. 258-259 '. By decomposition in a similar manner of the acid bruAnal. Calcd. for C14He04Cl4: C1, 37.4; neut. equiv., cine salt described above the d-acid was obtained, m. p. 190. Found: C1, 37.1, 37.1; neut. equiv., 189, 189. 252-254'. Rotation. 0.0349 g. made up to 1.95 cc. in chloroform a t The recrystallization from hot concentrated sulfuric acid confirms the statement of Huntress, Cliff and A t k i n ~ o n ~25' ~ gave a D +2.39; 1, 1; [ a ] z 6 D +133". that diphenic acids containing chlorine atoms in the 6,6'Anal. Calcd. for C14HaOaC14: C1, 37.4. Found: C1, positions do not undergo fluorenone formation. 36.6. I n some runs the crude was purified by heating it at 160" Examination of By-products.-From the 9 g. of crude deand 10 mm. in a sublimation apparatus. The non-sub- scribed above there was obtained 4.5 g. of yellow bylimable portion was then recrystallized from ethyl acetateproducts using either the sublimation procedure, where ligroin with use of decolorizing carbon. The product was they were obtained as the sublimate, or the sulfuric acid identical with that obtained above. The yellow sublimate procedure, where they were obtained after pouring the filis described below. trate and washings on crushed ice (water being added to Resolution of d,L4,6,4',6'-Tetrachlorodiphenic Acid.facilitate complete precipitation). The by-products seemed The optical resolution was carried out chiefly as a proof of identical with the exception that those obtained by the structure for the substance. All rotations were taken in a sulfuric acid method contained a small amount of non1-dm. semi-micro tube. acidic material which after recrystallization from dioxanewater consisted of light yellow needles, m. p. 244-249", (14) Morton, Znd. Eng. Chcm., Anal. Ed., 11, 170 (1939).
July, 1940
DIRECT SYNTHESISOF RESOLVABLE BIARYLS
containing 3.9% nitrogen and 38.7% chlorine. This material showed a strong green fluorescence in concentrated sulfuric acid or in alcohol. This fluorescence was quenched by the addition of acid to the alcohol solution. The aqueous dioxane filtrate gradually decomposed as it evaporated, giving a brown tarry solid with a charred odor. Neither this nor the yellow crystalline material was examined further. After elimination of this non-acidic material the byproducts were recrystallized from benzenealcohol. The first fraction consisted of pale yellow needles, m. p. 230231 weight 1.2 g.; this did not depress the melting point of authentic 3,5-dichloro-2-aminobenzoicacid. An additional 1.1 g. of material of m. p. 219-229' was obtained by concentration of the filtrate. The residues melted at still lower temperatures and were not examined. They probably contained 2,3,5-trichlorobenzoic acid for this substance had been isolated in earlier runs where the crude was leached with boiling water; it crystallized from the hot water in white needles, m. p. 161-163'. Cohen and Dakin16record m. p. 162". Anal. Calcd. for C7H302C18: C1, 47.3. Found: C1, 46.3. The 2.3 g. of material recovered actually contained 2 g. of 3,5-dichloro-2-aminobenzoic acid; this was shown by using the material along with similar material from other runs as starting material for a normal coupling run. Only 10% failed to diazotize and the yield of tetrachlorodiphenic acid was normal. If correction is made for the 2 g. of starting material recovered in usable form, the net yield of tetrachlorodiphenic acid is raised to 61%. The 3,5-dichloro-2-aminobenzoicacid recovered seems to be a true by-product; it does not result from incomplete diazotization. First, it is insoluble in acids of a concentration existing in the diazonium solution and the latter had been a t e r e d before use. Second, a quantitative analysis of a typical diazonium solution prior to use, based on the coupling reaction with &naphthol and using the salt pile indicator technique, showed complete diazotization. 3,5-Dibromo-2-aminobenzoicAcid.-Twenty grams of anthranilic acid was brominated following the procedure of Wheeler and Oates's with minor variations. The crude acid was obtained in yields of about 80% and was suitable for conversion to the tetrabromodiphenic acid. By recrystallization from alcohol-water it was obtained in 60% yields, m. p. 232-233 '. Wheeler and Oatesle record 232". Anal. Calcd. for C7H60zNBr2: Br, 54.2; N, 4.75; neut. equiv., 295. Found: Br, 54.4; N, 4.93; neut. equiv., 296. d,Z-4,6,4',6'-Tetrabromodiphenic Acid.-Ten grams (0.034 mole) of 3,5-dibromo-2-aminobenzoicacid and 3 g. (0.042 mole) of sodium nitrite were dissolved in 185cc. of water and 20 cc. of 2 N sodium hydroxide (0.04 mole). The diazotization was carried out as for the analogous chloro acid using 30 cc. of 6 N hydrochloric acid and 20 cc. of water. The reducing agent was prepared from 17 g. of purified cuprous chloride. The coupling suspension contained 24 cc. of ammonium hydroxide (28%) and 36 cc. of water. During the reaction a n additional 12 cc. of amO,
(15) Cohen and Dakin, J . Chem. Soc., 81, 1331 (1902). (16) Wheeler and Oates, THIS JOURNAL, 82, 770 (1910).
1707
monium hydroxide and 10 cc. of water was added. The manipulative technique was the same as that described above for the chloro acid. The crude was purified by either the sublimation or concentrated sulfuric acid procedures. Using the latter there was obtained 3.5 g. (37% yield) of the white crystalline product, m. p. 303-306 Recrystallization from ethyl acetate-ligroin gave a slightly purer product, m. p. 305308". Anal. Calcd. for C14HeO4Br4: Br, 57.3; neut. equiv., 279. Found: Br, 57.1, 57.2; neut. equiv., 279, 279. The by-products obtained by either purification procedure were treated as described for the chloro acid. From these in the run just described there was obtained 1.2 g. of pure 3,5-dibromo-2-aminobenzoic acid, m. p. 230-233 '. Moreover, the total amount of such material recovered in usable form was 3 g. The net yield of tetrabromodiphenic acid is then raised to 53%. Resolution of d,l-4,6,4',6'-Tetrabromodiphenic Acid.Five grams (0,009 mole) of the acid in 70 cc. of hot ethyl acetate was added quickly to a solution of 8 g. (0.0203 mole) of anhydrous brucine in 200 cc. of boiling ethyl acetate. In a few minutes 6.2 g. of salt separated from the boiling solution. This less soluble salt (2A.ZB) was leached with 400 cc. of boiling anhydrous ethyl alcohol. The residue and the first solid obtained on cooling the alcohol had about the same rotation and m. p. 259-260" decompn. Rotation, 0.0350 g. made up t o 1.95 cc. chloroform a t ~ %'gave CYD-0.19; I , 1; [ a I z 4-10.6'. Anal. Calcd. for C14H~~OaBrr.(CzsHze04Na)a:N, 4.16; C1, 23.7. Found: N, 4.54; C1, 23.2. The more soluble salt (dA.IB) was obtained only when the original ethyl acetate solution had been concentrated to 50 cc.; 1.8 g. so obtained was crystalline but even after drying for five hours a t 80" and 5 mm. it had m. p. 123204 O decompn. Rotation. 0.0333 g. made up to 1.95 cc. in chloroform at 25" gave CYD-0.55; I , 1; [ c Y ] ~ ~-32.2". D Anal. Calcd. for ClrHeO4Brc.(C13Ha~O~N~a)l: N, 4.16; Found: N, 4.10. There was also a considerable amount of glassy residue which when decomposed gave the d-acid with a rotation slightly greater than that given by the acid from the crystalline fraction of the more soluble salt. 1-4,6,4',6'-Tetrabromodiphenic Acid.-The less soluble salt was decomposed with dilute sodium hydroxide giving an acid which upon recrystallization from ethyl acetateligroin had m. p. 282-283'. Rotation. 0.1319 g. made up to 1.95 cc. in anhydrous ethyl alcohol" a t 25' gave CYD-0.52; I , 1; [aI2b-7.7". Anal. Calcd. for C14H60~Br4:Br, 57.3. Found: Br, 57.1. d-4,6,4',6'-Tetrabromodiphenic Acid.-The glassy residue was decomposed as above to give an acid which after recrystallization from ethyl acetate-ligroin had m. p. 279-282 '. Rotation. 0.0583 g. made up to 1.95 cc. in anhydrous D ethyl alcohol a t 25" gave a~ +0.20; 1, 1; [ C Y ] ~ ~+6.7".
'.
(17) This solvent was used for the bromo acids because of their very low solubility in chloroform.
M. C. KLOETZEL
1708
Anal. Calcd. for CI4HsO4Bra:Br, 57.3. Found: Br, 57.4. Racemization Experiments.-A sample of 1-4,6,4',6'tetrachlorodiphenic acid having [a]D - 129' (in chloroform) was refluxed in a 1 N sodium hydroxide solution for six hours. It was recovered by acidification and dried. It then showed [ a ] -126". ~ The decrease is probably not significant. A sample of d-4,6,4',6'-tetrabromodiphenicacid having [ a ] +6.7' ~ (in alcohol) was dissolved in a 1 N sodium hydroxide solution and kept a t 80" for five days. It was recovered by acidification and had the same rotation.
Summary 1. The biaryl synthesis involving the action
Vol. 62
of reducing agents on aqueous solutions of diazotized amines has been applied to the synthesis of resolvable diphenic acids. The method constitutes a rapid and direct synthesis of these substances. 2 . d,l-4,6,4',6'-Tetrachlorodiphenic acid and the analogous tetrabromodiphenic acid have been prepared by this method in one step from the appropriate 3,5-dihalogeno-2-aminobenzoic acids. Net yields obtained were 61 and 33%, respectively. The acids were resolved into their active forms by means of brucine. DURHAM, N. H.
RECEIVED APRIL 24, 1940
[CONTRIBUTION FROM THE CHEMICAL LaBORATORY O F HARVARD UNIVERSITY]
Polymethyl Aromatic Hydrocarbons. I. Synthesis of 1,2,4-Trirnethylnaphthalene and 1,2-,1,3- and 1,4-Dimethylnaphthalene BY M.C.KLOETZEL Research now in progress required the prepa- phenylpentanoyl chloride with aluminum chloride ration of certain di- and tri-methylnaphthalenes has been reported112to give (IV) in yields of 70in sizable quantity and free from isomers. 75%, the simpler direct cyclization of the acid Herein is described the synthesis of 1,2,4-tri- (111) with 80% sulfuric acid was employed and methylnaphthalene (XVI) and 1,2-, 1,3- and 1,4- the desired cyclic ketone (IV) was obtained dimethylnaphthalene by new methods which ful- thereby in 74% yield. fil these requirements. The readily available P-benzoylpropionic acid was utilized as the starting material for the preparation of 4-methyl-1-tetralone (IV) and from this cyclic ketone in turn were synthesized 1,3- and 1,4-dimethylnaphthalene and 1,2,4-trimethylnaphthalene (XVI). When methyl P-benzoylpropionate (I) was treated with methylmagnesium iodide 4-phenylpenten-3-oic acid (11) was produced. This acid has been prepared preR R viously by Mayer and Stamm' but in only 26% I I yield. I t was therefore of primary importance that this step in the synthetic scheme be improved. Consistent 70-75y0 yields of 4-phenylpenten-3-oic acid (11) finally were afforded by means of strict adherence to certain precautions, (VI, R = CHs) (V, R = CHs) (XVIII, R = H) (XVII, R = H) notably the use of 1.38moles of methylmagnesium R R iodide per mole of methyl P-benzoylpropionate, and maintenance. of the reaction temperature near 0' during addition of the Grignard reagent. An acetic acid solution of (11) rapidly absorbed hydrogen in the presence of Adams catalyst to 0 0 give a quantitative yield of 4-phenylpentanoic (VII, R = CHI) (VIII, R CHa) (XX, R = H) (XIX, R = H I acid (111). Inasmuch as the cyclization of 4- _____ (1) Mayer and Stamm, B e r . . 66, 1424 (1923).
(2) Von Rraun and Stuckenschmidt, ibrd., 66, 1724 (1923).
